Spotlight: Exposing the evolutionary past of chromosome 15

Mike Zody
Mike Zody

Making copies is what DNA does best. Its paired helical structure is specifically designed to create perfect copies of itself, forming a kind of molecular printing press for our genetic information. Sometimes, though, the press malfunctions, causing a small piece of DNA to be put back in the wrong place, subsequent to its duplication. Surprisingly, about 5% of our genetic makeup consists of such misplaced copies, termed "segmental duplications," which was one of the many tantalizing discoveries to emerge from the sequencing of the human genome. With the exception of other primates, this phenomenon appears to be more prevalent in humans than in any other mammal.

Chromosome 15 offers a particularly vivid illustration of this point. Though it accounts for about 2.9% of the total length of the human genome and contains about the same proportion of genes, it has far more than its share of segmental duplications. Most of these are unique to the chromosome and do not appear anywhere else in the genome. Together with my colleagues in the Broad's Genome Biology Program, I am working to determine why and how these segmental duplications arose in the first place. Our efforts could change the way we approach such duplicated regions in this part of the genome, as well as others, and may also contribute to our understanding of human genetic diversity and disease.

The duplicated sequences on chromosome 15 can be parsed into two distinct groups, which are physically divided by a long intervening region, largely devoid of such duplications. Curiously, almost all of the duplications are related, as we determined from their similar DNA sequences or their overlapping positions. This suggests that they have arisen solely from localized episodes of "copying and pasting" within the chromosome. But, if this is the case, how did these two separate, and seemingly independent, clusters emerge?

We looked for the answer using comparative genomics. In dogs and rodents, we discovered that the two regions of duplication on chromosome 15 actually sit side by side, flanking one another. In fact, they are likely to be positioned this way even in the mammalian ancestor. Therefore, we propose that the modern structure of chromosome 15 that we see in humans resulted from a genetic inversion, which turned one end of the chromosome over the other.

To accommodate such a maneuver, though, the chromosome must first break internally. In this case, the breakpoint must have been near, or within, the initial duplications, splitting the region in two. The subsequent inversion served to wedge the long, unduplicated segment of DNA between the two halves of the original duplicated locus, physically isolating them from each other. We predict that this occurred early in primate evolution and, while the individual clusters have continued to be copied and to expand locally, there has been no interchange between them since the divergence of our human ancestors from Old World monkeys.

We next wondered how the location of the chromosomal breakage that enabled the inversion was determined. Was it merely a coincidence that it occurred within the duplicated region or, could the duplications have somehow caused it? Preliminary evidence points to the latter. Chromosome 15 actually contains many breakpoints, which represent sites of genetic rearrangement that are specific to humans. Almost all of these show a significant enrichment for segmental duplications, demonstrating a possible causative link between duplication and rearrangement, though the exact nature of the association is still unclear.

What we can say with some certainty is that this sort of rearrangement isn't simply ancient history. Of the ten remaining gaps in the sequence of chromosome 15, seven are associated with segmental duplications. This is not surprising given that aligning the DNA sequences in duplicated regions is very difficult, like assembling a jigsaw puzzle whose pieces all have nearly identical shapes.

What is surprising, though, is that six of these seven regions have been reported to be potential sites of DNA copy number variation, a type of genetic polymorphism among individuals. In fact, in our attempts to bridge one of the sequence gaps in chromosome 15, we found that we could close it in two different ways: the linking pieces of DNA differed in length by approximately 100,000 nucleotides. The discrepant sequence harbored a duplicated chunk—that is, one piece had a segmental duplication and the other did not—and suggests that we indeed captured a region with copy number variation. If the other gaps that remain in the sequence of the human genome show similar variation—our data suggest that at least some do—it could explain the difficulties that have been encountered thus far in closing them.

Our work suggests that duplications, which sculpted human chromosome 15 from its predecessors, continue to shape its modern-day form, and thus contribute to the spectrum of human genetic diversity. Future studies will help us determine how much variability may lie undiscovered amidst these duplications and how this may relate to differences in disease susceptibility within the human population.

Paper(s) cited